Skip to Main Content
The main obstacles in current distributed hydrological modeling are the lack of sufficient data for driving the models and for parameterization of the land surface and subsurface. This study applied remote sensing (RS) based input data in a hydrological model for the 19012 km2 Kaidu River basin in Northwest China. Based on the geography information system (GIS) technique, the digital elevation model (DEM) of the study basin was successfully used to delineate the stream network and extract information of catchment characteristics. Landsat TM data from 1990 to 2001 have been utilized for various scale depended maps of land use and vegetation, and leaf area index (LAI) from MODIS data is used as input data to the distributed hydrological modeling. Further, root depths of annual vegetation are related to the temporal and spatial variation of LAI.The distributed hydrological model MIKE SHE was calibrated and validated against observed discharge for tow individual gauges during the period from 1998 to 2001. The model generally performed well for effective coefficient (R2), water balance error (RE) and correlation coefficient (r). It is shown that a well-calibrated MIKE SHE model with five "free" parameters is able to produce consistent results with correlation coefficients greater than 0.75. The potential for driving large scale hydrological models using remote sensing data was clearly demonstrated and further emphasized by the presence of long time records and near real time accessibility of the satellite data sources.
Computer Science and Information Engineering, 2009 WRI World Congress on (Volume:4 )
Date of Conference: March 31 2009-April 2 2009